microsoft/onnxruntime

Pull Requests Commits

Merge origin/main for PR #32665 readiness

tianleiwu committed 3 days ago

0a837973

Fix model package EP setup and stable API contracts (#32790)

xiaoyu-work committed 3 days ago

Verified 58bacbc1

[WebGPU EP] fixes fp16 accumulation bug in gemm kernel (#33057)

prathikr committed 3 days ago

Verified 0066c380

Add FP16 and BF16 H512 GQA decode support (#33176)

hanbitmyths committed 3 days ago

Verified 426d9b35

Fix empty CPU reductions with negative axes (#33232)

xadupre committed 4 days ago

Verified 9a6089e1

Fix PadFusion changing MaxPool results for negative inputs (#33206)

the0cp committed 4 days ago

Verified e19a9ae8

[WebGPU] Accelerated model loading using Win32 and D3D12 (#33055)

sushraja-msft committed 4 days ago

Verified f289d7a8

[CUDA] Fix QMoE SM120 configuration dispatch on SM121 devices (#33225)

kunal-vaishnavi committed 4 days ago

Verified a477effa

[WebGPU] Reduce QMoE single-token gate sorting to stable TopK (#33058)

sayanshaw24 committed 4 days ago

Verified 8fe86b17

Fix WASM Debug external-data test failures (#33226)

Copilot committed 4 days ago

Verified 2297e770

[CPU] Support mixed types and precision in QuantizeLinear (#32735)

TANGBUDU committed 4 days ago

Verified d7a4f6a3

Make shared WebGPU data transfers stateless (#33208)

mustjab committed 4 days ago

Verified ac44e162

[WebGPU] Guard padded workgroups after dispatch normalization (#33167)

edgchen1 committed 4 days ago

Verified 79d3fcf1

[WebGPU] Fix deferred pipeline callback lifetime after failed waits (#33209)

aciddelgado committed 4 days ago

Verified e379df24

Limit function call depth and fix lint issues (#32339)

apsonawane committed 4 days ago

Verified 17be9c28

Fix WebGPU nightly build (#33162)

edgchen1 committed 4 days ago

Verified 41c59c0f

[Compile API] Add support for writing external initializers and EP context to buffer to enable saving models > 2GB (#31347)

skottmckay committed 4 days ago

Verified 660125b3

Update parameterized requirement from >=0.8.1 to >=0.9.0 (#33022)

dependabot[bot] committed 4 days ago

Verified 06cab16f

fix(cuda): correct GatherBlockQuantized tests

tianleiwu committed 4 days ago

30b1f414

Fix isolated provider tests in Windows GPU CI (#33203)

xadupre committed 4 days ago

Verified 864d8712

Bump baseline-browser-mapping from 2.10.31 to 2.11.26 in /js/react_native/e2e (#32509)

dependabot[bot] committed 4 days ago

Verified 17447024

Fix complex STFT window handling (#33136)

xadupre committed 4 days ago

Verified be3b8e88

[CPU] Avoid degenerate GQA FlashAttention tiles when L2 cache size is unknown (#33202)

xadupre committed 4 days ago

Verified 40a7167d

Fix EP assignment reporting after NHWC layout transformation (#33184)

saiyambharara committed 4 days ago

Verified 430f563d

Improve GQA shared-cache and sliding-window handling (#33175)

hanbitmyths committed 4 days ago

Verified 26ab6c79

Improve CUDA RMSNorm reduction and aligned half2 output (#33178)

hanbitmyths committed 4 days ago

Verified d8024cdc

Fix WebGPU context memory identity and test host-owned caches (#33110)

mustjab committed 4 days ago

Verified 78cf3eb5

[WebGPU] Add 3D ConvTranspose support (#32688)

gyagp committed 4 days ago

Verified ceef7279

Bump source-map-js from 1.2.1 to 1.2.2 in /js/web/test/e2e/exports/testcases/nextjs-default (#33151)

dependabot[bot] committed 4 days ago

Verified c5776016

Use checked arithmetic for CPU LSTM state extents (#33128)

xadupre committed 5 days ago

Verified d8bd1bba

Older